Offline AI dictation has taken a dramatic leap forward with Google’s quiet release of a brand-new iOS app. Google AI Edge Eloquent landed in the App Store on April 6, 2026, with no press release and no formal announcement. Yet the app already poses a serious challenge to paid competitors like Wispr Flow, SuperWhisper, and Willow.
The tool converts raw, messy speech into polished, professional text. More importantly, this offline AI dictation experience works without sending a single byte of audio data to the cloud. As a result, users get both speed and privacy in one free package.
How Offline AI Dictation Works in Google Eloquent
Google built Eloquent on its Gemma-based automatic speech recognition models. Once a user downloads these models, all transcription runs directly on the device. In other words, the app needs no internet connection after the initial setup. That on-device architecture is what makes this offline AI dictation tool fundamentally different from cloud-dependent alternatives.
When recording begins, a live waveform tracks the speaker’s voice while text appears in real time. After the user hits pause, the system automatically strips filler words such as “um” and “ah.” It also smooths out mid-sentence corrections. Consequently, the final output reads closer to edited writing than a raw transcript.
This approach stands in sharp contrast to most competing tools. For instance, Otter.ai requires a constant internet connection for its core features. Meanwhile, Wispr Flow and Willow both charge around $15 per month for their premium tiers.
5 Features That Set Eloquent Apart from Paid Rivals
Several standout capabilities make this free app worth watching. First, Eloquent provides four text transformation modes. Users can convert their dictation into key-point summaries, formal prose, shortened versions, or expanded long-form text. As a result, a single recording session can produce multiple content formats.
Second, the app includes a personal context dictionary. Users can manually add names, technical jargon, or industry-specific terms to improve recognition accuracy. Furthermore, those who sign in with a Google account can import frequently used words from recent Gmail messages. This optional feature builds a vocabulary profile without any manual configuration.
Third, Eloquent tracks detailed usage metrics. The app records words-per-minute speed, total word count, and session history. Therefore, productivity-focused professionals can measure dictation output over time. The growing role of AI in finance and other regulated industries makes this kind of hands-free documentation increasingly relevant.
Fourth, the dual processing model offers genuine flexibility. In fully offline mode, nothing leaves the device. However, users who want deeper text polish can toggle on cloud mode, which sends text (not audio) to Google’s Gemini models for refinement. This layered architecture gives users direct control over their data while preserving the core offline AI dictation functionality.
Fifth, Google charges absolutely nothing. There are no subscriptions, no in-app purchases, and no usage caps. By comparison, most AI-powered startup tools in this space require monthly fees.
Why an iOS-First Launch Surprised the Tech Industry
Google releasing an app on Apple’s platform before its own Android ecosystem raised eyebrows across the industry. Typically, Google showcases new AI capabilities on Pixel hardware first. Nevertheless, the App Store listing references an upcoming Android version with features like default keyboard integration and a floating activation button.
According to TechCrunch, the app appeared without any accompanying blog post or media event. Chromeunboxed also reported that a “Download for Android” button briefly appeared on the app’s website before being pulled. This suggests the Android launch is imminent but was pushed live prematurely.
The move aligns with a broader industry shift toward on-device AI processing. Offline AI dictation fits squarely into this trend, as companies across sectors recognize that privacy-first technology resonates with both consumers and enterprise buyers. In turn, this trend is reshaping how businesses think about data handling and user trust.
What This Means for the Voice-to-Text Market
Eloquent’s arrival fundamentally changes the competitive landscape for offline AI dictation. Until now, users who wanted accurate, filler-free transcription had to pay for it. As a result, speech-to-text tools remained niche products for power users and professionals.
Google’s decision to offer comparable quality at zero cost could accelerate mainstream adoption. Similarly, it puts pressure on startups like Wispr Flow and Willow to differentiate on cross-platform support and enterprise features. The broader AI startup ecosystem is already adjusting to a world where large tech companies release competitive products for free.
For businesses in regulated industries, the fully offline mode carries additional appeal. Financial services firms, legal teams, and healthcare providers often restrict cloud-based tools due to compliance requirements. Consequently, a free, on-device offline AI dictation tool from a trusted brand fills a real gap in the market.
Whether Google continues supporting Eloquent or lets it languish as an experiment remains the open question. The company’s track record with iOS apps is mixed at best. Still, the technology works, the price is right, and the timing could not be better for an offline AI dictation tool that turns spoken words into clean, professional text without ever touching the cloud.